Build, Debug, Practice Modes
Ship with an agent when it matters. Learn to code when it doesn't.
Three explicit modes for working with a coding agent: BUILD delivers, DEBUG fixes or teaches, PRACTICE lets you write the code with five escalating hint levels, plus a Code Red modifier that never skips tests.
Why it's different
- Five hint levels, no solution before a real attempt
- Code Red skips teaching, never tests
- A weekly learning rhythm that works
Try asking
“learn to code while building”“guided debugging”“coding exercises with AI”
Details
- Works with
- Claude.ai, Claude Desktop, Claude Code, Codex and other SKILL.md agents
- Needs
- None
- Pack
- Claude Code & Codex Workflow (€29)
- Also covers
- learn to code with AI, vibe coding, pair programming with claude
Same pack
Also in Claude Code & Codex Workflow
Multi-Agent Repo Setup
Claude Code and Codex on one project, without stepping on each other.
gitView →
Overnight Agent Report
Wake up to a report you can act on in five minutes.
No setup neededView →
Manual Acceptance Checklist
The honest test list for what your agent couldn't check.
No setup neededView →
LLM Feature Guardrails
Make sure 'asdf' never reaches your model, and your keys never reach the browser.
No setup neededView →
Deploy Exposure Check
A private repo doesn't make your files private. Check what your site actually serves.
Python 3, gitView →